gmakstutis/cognee
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fixes to delete

Browse source
This commit is contained in:
vasilije 2025-07-19 19:06:10 +02:00
parent 9d423f5e16
commit bf191ae6d0
1 changed files with 19 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

22
cognee/modules/ingestion/data_types/TextData.py
View file

@ -16,9 +16,12 @@ class TextData(IngestionData):
self.data = data self.data = data
def get_identifier(self): def get_identifier(self):
keywords = extract_keywords(self.data) import hashlib
return "text/plain" + "_" + "|".join(keywords) content_bytes = self.data.encode("utf-8")
content_hash = hashlib.md5(content_bytes).hexdigest()
return "text/plain" + "_" + content_hash
def get_metadata(self): def get_metadata(self):
self.ensure_metadata() self.ensure_metadata()
@ -27,7 +30,20 @@ class TextData(IngestionData):
def ensure_metadata(self): def ensure_metadata(self):
if self.metadata is None: if self.metadata is None:
self.metadata = {} import hashlib
keywords = extract_keywords(self.data)
content_bytes = self.data.encode("utf-8")
content_hash = hashlib.md5(content_bytes).hexdigest()
self.metadata = {
"keywords": keywords,
"content_hash": content_hash,
"content_type": "text/plain",
"mime_type": "text/plain",
"extension": "txt",
"file_size": len(content_bytes),
}
@asynccontextmanager @asynccontextmanager
async def get_data(self): async def get_data(self):